Summary:
There was a fix around folly::dynamic constructor that will be needed for Fabric work. This was done in 94e964976c (diff-7d1cb97d222ba0c863ea8a8e43b2ee2b) and luckily the release 1 month after the Folly version we used in RN already had the fix, so that we don't need to upgrade to the latest folly yet (minimizing breakages).
Tested by:
* running RNTester xcode project (ios)
* running RNTesterPods workspace via cocoapods
* building android via gradle
Reviewed By: shergin
Differential Revision: D7626037
fbshipit-source-id: cb36ba5b91ba131d4e450300bd620db657cfa1e8
Summary: Should have been checked in previously. No setup change here.
Reviewed By: shergin
Differential Revision: D7626040
fbshipit-source-id: e7a1a9de70a40a22e39eb3777d1b62c8d7b66381
Summary:
This adds a way to test out CocoaPods build to RNTester, but as a separate project/workspace.
This also fixes the podspecs due to Fabric stuffs.
Note that this setup is meant to be run manually, not by CI (yet)
Reviewed By: mmmulani
Differential Revision: D7603823
fbshipit-source-id: 2219aa4c77f40ec07097a5c9ed4052529226618b
Summary:
On Android, LayoutAnimation directly updates the layout since a generic
scaling animation is more difficult to implement. This causes a problem
if the layout is updated during an animation, as the previous layout is
stored with the animation and is not updated. As a result the view gets
the old layout instead once the animation completes.
This commit fixes this issue by storing the layout handling animations
while those animations are active, and updating the animations on the
fly if one of the views receives a new layout. The resulting behaviour
mirrors what iOS currently does.
This bug has real world consequences, for example if a LayoutAnimation
happens right after a VirtualizedList has mounted, it’s possible that
some list rows are mounted while the animation is active, making the
list content view bigger. If the content view is being animated, the
new size will not take effect and it becomes impossible to scroll to
the end of the list.
I wrote a minimal test case to verify the bug, which I’ve also added to
RNTester. You can find the standalone app here:
<https://gist.github.com/lnikkila/18096c15b2fb99b232795ef59f8fb0cd>
The app creates a 100x300 view that gets animated to 200x300 using
LayoutAnimation. In the middle of that animation, the view’s dimensions
are updated to 300x300.
The expected result (which is currently exhibited by iOS) is that the
view’s dimensions after the animation would be 300x300. On Android the
view keeps the 200x300 dimensions since the animation overrides the
layout update.
The test app could probably be turned into an integration test by
measuring the view through UIManager after the animation, however I
don’t have time to do that right now...
Here are some GIFs to compare, click to expand:
<details>
<summary><b>Current master (iOS vs Android)</b></summary>
<p></p>
<img src="https://user-images.githubusercontent.com/1291143/38191325-f1aeb3d4-3670-11e8-8aca-14e7b24e2946.gif" height="400" /><img src="https://user-images.githubusercontent.com/1291143/38191337-f643fd8c-3670-11e8-9aac-531a32cc0a67.gif" height="400" />
</details><p></p>
<details>
<summary><b>With this patch (iOS vs Android, fixed)</b></summary>
<p></p>
<img src="https://user-images.githubusercontent.com/1291143/38191325-f1aeb3d4-3670-11e8-8aca-14e7b24e2946.gif" height="400" /><img src="https://user-images.githubusercontent.com/1291143/38191355-07f6e972-3671-11e8-8ad2-130d06d0d64d.gif" height="400" />
</details><p></p>
No documentation changes needed.
[ANDROID] [BUGFIX] [LayoutAnimation] - View layout is updated correctly during an ongoing LayoutAnimation, mirroring iOS behaviour.
Closes https://github.com/facebook/react-native/pull/18651
Differential Revision: D7604698
Pulled By: hramos
fbshipit-source-id: 4d114682fd540419b7447e999910e05726f42b39
Summary:
BREAKING CHANGE
This change upgrades the React Native build pipeline from Babel 6 to Babel 7
If you use a `.babelrc` then you'll need to update it to Babel 7 (note that some plugins are no longer relevant, some plugins are automatically upgraded, and some will need some manual love).
Note that you may also need to upgrade your dev env, tests etc, to make sure they work with Babel 7.
Reviewed By: mjesun
Differential Revision: D7591303
fbshipit-source-id: 29cef21f6466633a9c366d1f3c0d3cf874c714db
Summary:
BREAKING CHANGE
This change upgrades the React Native build pipeline from Babel 6 to Babel 7
If you use a `.babelrc` then you'll need to update it to Babel 7 (note that some plugins are no longer relevant, some plugins are automatically upgraded, and some will need some manual love).
Note that you may also need to upgrade your dev env, tests etc, to make sure they work with Babel 7.
Reviewed By: cpojer
Differential Revision: D7097279
fbshipit-source-id: 9fb204cae733174a1c155669b7c17ddb70f7aecc
Summary:
The latest release of react-native (0.55.2) does not expose the new `InputAccessoryView` component; It can't be accessed at all. This change fixes this problem.
* Problem: Snack showing the problem: https://snack.expo.io/B1fDQRYif
* Proof: `RNTester` still works with adapted imports
No related PRs.
[IOS] [BUGFIX] [InputAccessoryView] - Expose `InputAccessoryView` so it can be imported
Closes https://github.com/facebook/react-native/pull/18780
Differential Revision: D7581729
Pulled By: hramos
fbshipit-source-id: d61ab1f167360e829e32b93fb5414d2f7e57e115
Summary:
Apparently, there is no point to return a reference from this method because the struct (std::shared_ptr) is allocated on stack.
The test is also updated.
allow-large-files
Reviewed By: fkgozali
Differential Revision: D7557749
fbshipit-source-id: aa74146322c6d340256752586f05fc672024038e
Summary:
The previous name conflicts with the method with same (but with different semantic) name in `ShadowNode` class.
That was bad idea to use same name especially because the different semantic.
Reviewed By: fkgozali
Differential Revision: D7554549
fbshipit-source-id: 0bccbaacd0812f8a26592b2008f15ddb5bc34ebc
Summary:
`enum class` types do not provide default conversion to integers and reguire use typename before value names.
This must prevent bugs like the previous one where we used `Undefined` as a number value.
Reviewed By: fkgozali
Differential Revision: D7554548
fbshipit-source-id: b19379aae48c9aebe03043e08cf3acc712da3cb8
Summary: `Undefined` is actually equal `0` (becasuse it is `LayoutDirection::Undefined`), whereas `kFloatUndefined` is a magic huge number.
Reviewed By: fkgozali
Differential Revision: D7554550
fbshipit-source-id: 3ea37f0b8b6a59b4b0e00d205b75cd7252247d03
Summary:
Implementation of the native view component of <View>.
Root View also must have dedicated class even if its empty now.
Reviewed By: mdvacca
Differential Revision: D7526406
fbshipit-source-id: 9e14c9f679396c2cdb2d64c99086bb19df37c25c
Summary: Quite obvious. We have to check `sealable` flag of the child (not a parent) before mutating it.
Reviewed By: mdvacca
Differential Revision: D7526407
fbshipit-source-id: ce8e0d6446ff7eb23baee9c92f246d5e198fe377
Summary:
RCTSurfacePresenter coordinates presenting of React Native Surfaces and represents application-facing interface of running React Native core.
SurfacePresenter incapsulates a bridge object inside and discourages direct access to it.
Reviewed By: mdvacca
Differential Revision: D7526413
fbshipit-source-id: 9b6c513ec63a84d8b6c0951458cbd6b47f49669b
Summary:
RCTScheduler represent facebook::react::Scheduler as a Obejctive-C object.
It supposed to be single, unified, bidirectional interop layer between C++ and Obejctive-C worlds.
Reviewed By: mdvacca
Differential Revision: D7526405
fbshipit-source-id: a206755f64f2904981b356a40e7659922b24d7bb
Summary: RCTFabricSurface was reimplemented as separate class which does not rely on (old) RCTUIManager and delegate some functionality (which must be coordinated between Surface instances and Scheduler) to RCTSurfacePresenter.
Reviewed By: mdvacca
Differential Revision: D7526404
fbshipit-source-id: e8c38261bc489fd6066ba29a829cd8f623c26217
Summary:
YogaLayoutableShadowNode::yogaNodeCloneCallbackConnector recently was disabled because of change in Yoga (D7339832).
This diff brings is back.
Reviewed By: mdvacca
Differential Revision: D7526417
fbshipit-source-id: 5369d641bf1e118132cf742d2d243bf426c0ffdb
Summary:
A registry for components. The registry allows to refer to component view instance by react tag.
It also has recycle-pool-like interface (and eventually the pool will be implemented).
Reviewed By: mdvacca
Differential Revision: D7507519
fbshipit-source-id: ed74203fb4d0694490b0dc23aae0aa067b3fc55b
Summary:
MountItem is a small granular light-weight instruction describing a change in the component view tree.
All instruction are supposed to be small and standard. All additional logic must be implemented in component view subclasses.
No more opaque and untyped objcblocks.
All that allows mounting manager to control and optimize an execution of those items (eventually).
Besides that five we probably will need `perform imperative instruction` (for execution something like `scrollTo`) and `update local data` (for updating local state of something like text input).
Reviewed By: mdvacca
Differential Revision: D7507518
fbshipit-source-id: 745b93125231a02cbc152cfa6c6baf423d100f81
Summary:
All methods of RCTComponentViewProtocol are non-optional,
but this default implementation removes necessity to implement all of them manually.
Reviewed By: mdvacca
Differential Revision: D7507522
fbshipit-source-id: e4dff97e8bf64e79d2f1ffca94f0549bd4b5e2fa
Summary: RCTComponentViewProtocol is a protocol for all UIView instances which are managed by React/MountingManager.
Reviewed By: mdvacca
Differential Revision: D7507520
fbshipit-source-id: dcbed8c5d0cc97f72942508aeccb445833ade271
Summary: UIManager uses UIManagerDelegate to communicate about shadow tree changes to another parts of the system.
Reviewed By: fkgozali
Differential Revision: D7503484
fbshipit-source-id: 0afe0f0d6cad31fe2ee9d61235d02b379cfe8217
Summary:
We have to call shallowSourceNode() in all cases of cloning which were not caused by UIManager instructions,
otherwise the diffing alogorith might produce incorrect mutation instructions.
Reviewed By: mdvacca
Differential Revision: D7503383
fbshipit-source-id: b33e5c39b7ba8cbd0f925fd29b3af379441a40a4
Summary: We have to call `ensureunSealed()` only if the node was changed.
Reviewed By: mdvacca
Differential Revision: D7503388
fbshipit-source-id: a3d07d50fa983ef93c14fa771711fa783fdf4c12
Summary:
Previously we generated `removed` *or* `delete` instruction, but sometimes we have to generate both.
So, basically we did it wrong. :(
Reviewed By: mdvacca
Differential Revision: D7503386
fbshipit-source-id: 8ee476abd29f088f31dc776f6e6a68d5293fbb35
Summary:
The method replaces the current source node with its source node.
It's useful when we have to clone some node but don't want to change a source node pointer.
Reviewed By: fkgozali
Differential Revision: D7503384
fbshipit-source-id: 81ec64079c7e99cb9abdda2af10d85281a94e1b1
Summary:
Quite trivial.
Note that std::unordered_map's `operator[]` is not `const`, so we have to use `at` instead.
Reviewed By: mdvacca
Differential Revision: D7467799
fbshipit-source-id: df38b21dccee4b347f7c070600af0d52f38d6570
Summary:
The first and quite naive implementation of The Diffing algorithm.
The exact set of instructions, their semantic, order, amount, and excessiveness are still unclear.
The concept should be verified by comprehensive testing with working native views rendering layer.
Reviewed By: mdvacca
Differential Revision: D7467790
fbshipit-source-id: 08f2f646e058cac8a4b73bf7b148e2748633348d
Summary:
Motivation:
* We never should call `markDirtyAndPropogate()` during tree construction/mutation because it might affect trees in different thread/dimentions;
* In Fabric we basically always have to dirty nodes ourselves manually duting tree construction;
* In Fabric we don't have "scoped/limited" tree mutations which require recursive dirtying; any mutation is creation of the new tree instance;
* Default value of the `isDirty` flag is "false", so we have to change this right after creation of Yoga node (and after cloning).
Reviewed By: mdvacca
Differential Revision: D7467797
fbshipit-source-id: 2c9144271dceea6ba2b95173209b99b5d86fbd87
Summary:
The modern Concurent Yoga's concept is:
We have to set parent/owner reference as part of `appendChild` process only if the current reference to parent/owner is `null`.
The motivation:
* Null-parent indicates that this node was not attached to anything yet;
* So, in this case there is no any concurrent memory access because we always create and (first time) attach the node on same thread;
* Simmetrical parent-child relationship indicates that we don't need to (re)clone assosiated ShadowNode (nor Yoga node).
Reviewed By: mdvacca
Differential Revision: D7467791
fbshipit-source-id: 9a7f517380fde3bb00272de18fd5dc13edb52071
Summary: Previously we recreate a vector with pointers to child nodes every single time we modify the collection. That was okay but recently I realized that the we can simply make a copy of the vector one time during object construction and then mutate it freely.
Reviewed By: mdvacca
Differential Revision: D7467796
fbshipit-source-id: 660f1706a19ae5f07c34c509f411ce9d67b93b35
Summary:
Computed `layoutMetrics` are also considered as part of ViewShadowNode's value.
In the future we probably have to add something like `localData` and `imperativeCommands`.
We need all this for diffing algorithm and mointing phase.
Reviewed By: mdvacca
Differential Revision: D7467800
fbshipit-source-id: 8a0dcf1fd2f97dc501d6969cb0b0f6a2c6a648b4
Summary: Using methods of the base class instead of custom implementation.
Reviewed By: fkgozali
Differential Revision: D7467795
fbshipit-source-id: 4d168b72880f6900bf8b747e1d655c10140e0c79
Summary: We have to have getters for all props/fields.
Reviewed By: mdvacca
Differential Revision: D7467792
fbshipit-source-id: 1492aad2d3398e6c14e0e354047730cf91201175
Kevin Gozali
Himabindu Gadupudi
Peter van der Zee
Jakub Grzmiel
Leo Nikkilä
Aaron Brager
David Vacca
Sebastian Markbage
Rafael Oleza
Matt Mahoney
Manuel Alabor
David Aurelio
Valentin Shergin
